Research work entitled “Mutually Guided Image Filtering” [1] by Xiaojie Guo with the High Performance and Visual Computing Team (College of Intelligence and Computing at Tianjin University) collaborated with Dr. Yu Li (ADSC, Singapore), Prof. Jiayi Ma (Wuhan University), and Prof. Haibin Ling (Temple University, USA) has been recently accepted by the IEEE Transactions on Pattern Analysis and Machine Intelligence, one of the most famous and top-tier journals in the field of AI (impact factor 9.455).  

Image filtering is an important and fundamental component to many intelligent vision based systems, as well as a key step towards solving various problems/tasks in multimedia, computer vision and graphics, which aims to simultaneously suppress/eliminate unwanted information and preserve the desired one in images/video frames. The proposed Mutually Guided Image Filtering (muGIF) formulates multiple filtering modes in a unified framework, and customizes an effective and efficient algorithm to optimize the model. Its efficacy and superiority over other state-of-the-art alternatives have been demonstrated on a number of applications. 

Figure 1: Examples of muGIF in multiple filtering modes    

[1] Xiaojie Guo, Yu Li, Jiayi Ma, and Haibin Ling, Mutually Guided Image Filtering, Transactions on Pattern Analysis and Machine Intelligence (IEEE TPAMI) 2018. DOI:10.1109/TPAMI.2018.2883553 

Paper: https://ieeexplore.ieee.org/stamp/stamp.jsp?tp=&arnumber=8550683 

Project Page: https://sites.google.com/view/xjguo/mugif 

By: Guo Xiaojie, Li Na

Editors: Qin Mian and Keith Harrington